Our verdict
Solvd is a California-headquartered AI engineering firm with a decade-plus delivery history, 150+ AI and ML specialists, NeurIPS 2025 Best Paper recognition through its research leadership, and public case work for brands such as NerdWallet and Under Armour. It suits an enterprise that wants production AI, vision, or agentic software engineered and run at scale, and it is less suited to a buyer who wants a tiny specialist boutique.
The right buyer
- Enterprises that need production AI systems, not demos. It frames its work as building, implementing, and running AI where the stakes are real, with 750+ staff and 150+ AI/ML specialists (about page).
- Retail and marketplace teams improving onboarding or product understanding with AI. Published cases cut session cost 25.5% for a retail onboarding flow and beat a much larger vision baseline on product attributes (case study; case study).
- Buyers who want research-grade ML depth. Its chief scientist co-authored a NeurIPS 2025 Best Paper, and the Tooploox acquisition added AI research capacity (about page).
Reasons to look elsewhere
- At 750+ people it is larger than many boutiques, so a buyer who wants a five-person embedded team may feel oversized.
- Much of its engineering sits outside the US, so an onshore-only mandate needs an explicit staffing plan.
Solvd, At A High Level
Solvd was founded in Roseville, California, in 2011 around rigorous test automation, later expanded into digital engineering, and relaunched as an AI-first firm in 2025 after buying EastBanc Technologies in October 2024 and Tooploox in June 2025 (about page). Mike Hulbert is CEO; Igor Vayner is executive chairman; Alexey Khursevich, a co-founder, has served as CTO and later chief digital transformation officer (leadership page; Forbes Technology Council). Headquarters are at 1646 N. California Blvd. in Walnut Creek, with a Washington, DC, office and delivery centers in Poland, Ukraine, and Georgia (about page).
- 750+ global team members across 10 locations; we estimate annual revenue at $60 million to $90 million (about page).
- Clients and logos on its site include NerdWallet, Under Armour, Reddit, Nasdaq, and The Washington Post (Solvd).
- For a large retailer it replaced a seventy-question onboarding flow with an agentic conversation and cut average session cost 25.5% (case study).
- For marketplace product understanding, a fine-tuned 4B vision model beat a zero-shot 27B baseline on F1 while running about 3.8x faster (case study).
Scores by criterion
We scored Solvd from its about page, leadership page, and published AI case studies. Stability and technical depth score highest; commercial clarity scores lowest because prices are unpublished.
| Criterion | Weight | Score | Basis |
|---|---|---|---|
| Production track record | 20% | 7.5 | Public AI case results with quantified gains plus long-running enterprise QA and engineering work. |
| Technical depth | 15% | 7.5 | Core AI, agentic software engineering, and research-backed ML. |
| Industry depth | 10% | 7.0 | Finance, fintech, e-commerce, retail, media, logistics, and healthcare. |
| Delivery model and handoff | 15% | 7.0 | Build, implement, and run model with global delivery centers. |
| Commercial clarity | 10% | 5.5 | No published rates or minimums. |
| Platform independence | 10% | 7.0 | Works across major clouds and open models, without a single lock-in. |
| Stability and scale | 10% | 8.0 | Founded 2011, 750+ staff, two AI acquisitions in 2024-2025. |
| Independent evidence | 10% | 7.0 | Brand clients on its site and NeurIPS 2025 Best Paper credit. |
| Overall | 100% | 7.1 | Weighted average |
Where Solvd fits
Solvd is used by large consumer, fintech, and enterprise brands that want AI features engineered into production products and operations. On price it sits in the upper mid-market tier with a large global engineering base. Competitors on this site include BairesDev and Simform for large-scale product engineering with AI, and Dataroot Labs for applied ML.
| Need | Fit | Why |
|---|---|---|
| Production AI in a customer-facing product | Strong | Retail and marketplace AI cases are published with metrics. |
| Vision or LLM systems that must hit cost targets | Strong | Its vision and onboarding cases emphasize accuracy and unit cost. |
| A long-running engineering partnership | Good | Roots in QA and digital engineering since 2011. |
| A five-person specialist boutique | Weak | It operates as a 750-person global firm. |
What clients and leaders say
The buyer views below on Solvd are written from public facts about the firm and its work.
“They shipped the agentic onboarding flow into a real pilot with cost metrics we could take to finance, but the team was spread across several countries.”
“The vision model work was stronger on benchmarks than on turnkey productization, so we kept ownership of the last mile.”
“For a company our size they had enough AI specialists, though we had to ask early who would stay on after the research spike.”
Where customer reviews couldn’t be found, our analysts attempted to give a voice to buyer sentiment, drawing on real information from review sites, forums and product documentation. Based on: Solvd’s retail onboarding case study and its about page, which list ten global locations; Solvd’s product-photo vision case study, which notes productization sat outside the team’s scope; Solvd’s about page, which cites 150+ AI/ML specialists inside a 750+ firm.
Who leads Solvd
Mike Hulbert is Chief Executive Officer and a board member of Solvd (leadership page).
